|
Activity Number:
|
78
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 4, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #300192 |
|
Title:
|
Shrinkage Estimators for Robust and Efficient Inference in Haplotype-Based Case-Control Studies
|
|
Author(s):
|
Yi-Hau Chen*+ and Nilanjan Chatterjee and Raymond Carroll
|
|
Companies:
|
Institute of Statistical Science, Academia Sinica and National Cancer Institute and Texas A&M University
|
|
Address:
|
No 128 Sec 2 Academia Road, Taipei, 11529, Taiwan
|
|
Keywords:
|
bias-efficiency trade-off ; empirical Bayes ; gene-environment interaction ; penalized likelihood
|
|
Abstract:
|
We develop methods for haplotype-based analysis of case-control studies which gain efficiency by exploiting model assumptions of HWE and G-E independence and yet are resistant to bias when those assumptions are violated. The idea involves shrinkage of a "model-free" estimator that is robust to HWE and G-E independence towards a "model-based" estimator that directly exploits those assumptions. A "model-free" estimator is first proposed which is closely related to the "prospective" estimator. Considering the "retrospective" estimator in Spinka et al. (2005) as the "model-based estimator, we then propose shrinkage estimators based respectively on the empirical Bayes and the penalized likelihood approaches. Variance estimation for these shrinkage estimators is proposed. Bias vs. efficiency trade-off for the shrinkage estimators is studied via simulations, and a real application is provided.
|
